Recent advancements in deep learning-based remote photoplethysmography methods

Abstract

Health monitoring of an individual is guided by physiological parameters that can be estimated by a photoplethysmography signal using contact-based or contactless approaches. Contactless monitoring, being more advantageous than the contact-based approach, is an active area of research. Additionally, conventional contactless approaches are based on certain assumptions, which are not required for deep learning methods. Therefore, this chapter reviews the deep learning-based remote photoplethysmography signal extraction methods, making the following contributions: first, it presents various compressed and uncompressed datasets used in this domain; second, it summarizes and analyzes the region of interest selection methods, followed by remote photoplethysmography signal extraction methods based on the deep learning architecture baselines; finally, the limitations of the existing methods are highlighted with recommendations for future studies.